Best platform to monitor which prompts mention my brand: what we use

The first thing clients ask once they accept that AI search matters is simple: which questions are people asking where we show up? It is a fair question, and a surprisingly specific one. A single visibility percentage does not answer it. You need to see the prompts, the answers and the order brands appear in.

Here is how we set this up for client programs, and what we look for in a platform.

A score is not a prompt report

Many trackers reduce everything to one number. That is fine for a headline, but when a client asks "do we appear when someone asks for the best option for a team of fifty?", you need to open that exact prompt and read what each engine said. We care about four things.

The raw responses. Before trusting any aggregate, we read a sample of real answers. Sometimes the tool has counted a mention that is really a passing reference in a list. Sometimes it has missed one because the brand was named by a product name.

Rules for what counts. In Promptwatch, the Brand Book holds analysis instructions that decide what counts as a mention, and brand aliases catch spelling variants and product names. Setting this up properly is the first hour of any engagement. Skipping it produces numbers nobody believes.

Position as well as presence. Position in Answer separates being the recommendation from being item four in a list. Sentiment, scored from 1 to 100 per response, helps catch the cases where a brand is mentioned and described badly.

Competitors in the same frame. Competitor comparison prompt types, Share of Voice and a heatmap that compares any two dimensions let us show a client which rivals appear in the prompts they do not.

Building the prompt set

Most of the value comes from how prompts are written. We organize them into topics and tag them by prompt type: organic questions that never name the brand, brand questions, and competitor comparisons. Intent matters too, because someone researching a category behaves differently from someone choosing between two named options.

Two features save us time here. Prompt Explorer suggests prompts from Search Console, ranking keywords, People Also Ask, or any page we point it at. And query fan-outs show the sub-queries a model runs before answering, which often reveal angles the client never considered. Prompt Volumes add search volumes and difficulty scores, so we can tell the client which prompts are likely worth the effort.

Personas, with persona stacking, let us run the same question as a procurement manager and as a founder. The answers can differ, and so can the brands named.

Prompts run daily per model, country and language. That cadence matters, because answers vary from one run to the next, and we read trends over weeks rather than reacting to one data point.

What to do with a missing mention

Finding the prompts that do not mention a client is the start of the work. For each gap, we ask a short series of questions.

First, which pages are the engines citing instead? Citation analytics show page and domain level citations, and the Page Tracker lets us follow a competitor's URL. Second, can the engines reach the client's relevant pages at all? Agent Analytics shows IP-verified AI crawler visits, errors and the crawl-to-citation path. Third, is the client simply absent from the sources engines trust, such as Reddit threads or YouTube videos? The citation reports for those channels, plus offsite mentions, cover that.

If the cause is missing content, the content gap analysis and the Answer Gap Report frame what to write. Content generation tools then help turn gaps into drafts, which an editor still reviews.

Plans and fit

For a single brand, Essential at $95 a month includes 50 prompts and 6,000 responses. Professional at $245 has 150 prompts and 18,000 responses, plus crawler logs, which Essential does not have. For agencies, plans from $199 a month include unlimited prompts and projects. The free Explore tier is useful for a first look, but it is limited to ChatGPT and 10 prompts.

Caveats we share with clients

No monitoring platform can guarantee a recommendation. Tracked prompts are a sample of what people ask, not the whole of it. And a mention that the tool counts is a signal, not a sale.

With those limits understood, the setup above gives clients what they were really asking for: a list of the questions that matter, an honest view of where they appear, and a reason for each gap. For our programs, Promptwatch is where all of that sits together.